Research Article Cite This: ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
www.acsami.org
In Situ Characterization of Protein Corona Formation on Silica Microparticles Using Confocal Laser Scanning Microscopy Combined with Microfluidics Alessia C. G. Weiss,† Kilian Krüger,‡,§ Quinn A. Besford,† Mathias Schlenk,‡ Kristian Kempe,∥ Stephan Förster,‡,§,⊥ and Frank Caruso*,†
ACS Appl. Mater. Interfaces Downloaded from pubs.acs.org by EAST CAROLINA UNIV on 01/03/19. For personal use only.
†
ARC Centre of Excellence in Convergent Bio−Nano Science and Technology, and the Department of Chemical Engineering, The University of Melbourne, Parkville, 3010 Victoria, Australia ‡ Physical Chemistry I, University of Bayreuth, Universitätsstraβe 30, 95447 Bayreuth, Germany § JCSN-1/ICS-1, Forschungszentrum Jülich GmbH, Wilhelm-Johnen-Straβe, 52428 Jülich, Germany ∥ ARC Centre of Excellence in Convergent Bio−Nano Science and Technology, and Monash Institute of Pharmaceutical Sciences, Monash University, Parkville, 3052 Victoria, Australia ⊥ Physical Chemistry, RWTH Aachen University, 52074 Aachen, Germany S Supporting Information *
ABSTRACT: In biological fluids, proteins bind to particles, forming so-called protein coronas. Such adsorbed protein layers significantly influence the biological interactions of particles, both in vitro and in vivo. The adsorbed protein layer is generally described as a two-component system comprising “hard” and “soft” protein coronas. However, a comprehensive picture regarding the protein corona structure is lacking. Herein, we introduce an experimental approach that allows for in situ monitoring of protein adsorption onto silica microparticles. The technique, which mimics flow in vascularized tumors, combines confocal laser scanning microscopy with microfluidics and allows the study of the time-evolution of protein corona formation. Our results show that protein corona formation is kinetically divided into three different phases: phase 1, proteins irreversibly and directly bound (under physiologically relevant conditions) to the particle surface; phase 2, irreversibly bound proteins interacting with preadsorbed proteins, and phase 3, reversibly bound “soft” protein corona proteins. Additionally, we investigate particle−protein interactions on low-fouling zwitterionic-coated particles where the adsorption of irreversibly bound proteins does not occur, and on such particles, only a “soft” protein corona is formed. The reported approach offers the potential to define new state-of-the art procedures for kinetics and protein fouling experiments. KEYWORDS: kinetics, particles, nanoengineering, low-fouling, adsorption
■
INTRODUCTION Nano- and micro-sized particles are widely studied owing to their potential in drug delivery, therapy, and diagnostics. An important aspect in their development is to establish a fundamental understanding of protein adsorption onto themupon formation of the so-called protein corona, the particle “synthetic identity” transitions to a “biological identity”, subsequently influencing the physiological and therapeutic response of the particles.1−3 The protein corona is generally described as a two-component system.4 Proteins with a high affinity for the particle surface form a tightly bound layer, the “hard” protein corona. This layer is surrounded by a protein cloud, often referred to as the “soft” protein corona, wherein rapid dynamic exchange of proteins between the solution medium and particles dominates.5,6 However, because of the limited number of suitable characterization methods for monitoring and evaluating protein adsorption in detail, it is difficult to clearly define the different layers to confirm the general description used for protein coronas.7−9 Depending on © XXXX American Chemical Society
the characterization method used, the protein corona is described according to either the Gibbs free energy ΔG,8,10−12 which defines the adsorption and desorption rates of proteins, or binding force13,14 between the proteins and particle surface. Proteins with a large ΔG have a low probability of desorption and therefore remain associated with the particle surface. These proteins are considered to form the “hard” protein corona. Distinction based on binding forces implies that “hard” protein corona proteins interact directly with the particle surface through long-range, strong protein− surface interactions, whereas proteins in the “soft” protein corona interact with other proteins through short-range, weak protein−protein interactions. Another theoretical distinction is based on the persistence of the protein to remain adsorbed throughout the nanoparticle’s journey (i.e., from bloodstream Received: August 20, 2018 Accepted: December 11, 2018
A
DOI: 10.1021/acsami.8b14307 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
protein adsorption kinetics slow down with increasing fluid velocities.
to tissue and past-endocytic environments) as protein corona composition changes during biophysical events.6−8,15,16 The concept of “persistent” proteins originates from studies where the “hard” protein corona is used to follow the particle’s past.17−19 It is becoming increasingly important to clearly understand the complex process of protein corona formation, with a focus on the influence of the “soft” protein corona on physiological interactions.4,7,13,20−24 However, to do so, it is crucial to acquire and understand further details such as the time-evolution of protein corona formation. Existing techniques for investigating the protein corona can be divided into ex situ and in situ methods. Ex situ approaches study the protein corona after isolation of the particles from the biological environment.7 This process may change the protein corona composition and may not preserve the “soft” protein corona.8,25,26 Therefore, such techniques are essentially limited to the analysis of the “hard” protein corona. In contrast, in situ methods monitor the particle−protein complex directly in the incubation solution, allowing for analysis of the “soft” protein corona. Insights into the corona thickness, protein surface affinities, and stoichiometries of protein association with, and dissociation from, the nanoparticle surface, as well as the protein structure can be obtained. Commonly used analytical tools include dynamic light scattering,27,28 fluorescence correlation spectroscopy,23,25,29 zeta potential measurements,27,30 circular dichroism spectroscopy,31,32 Fourier transform infrared spectroscopy,32 and isothermal titration calorimetry.33 However, to our knowledge, there is no technique that enables the study of the kinetic formation of the entire protein corona in situ and under flow and further providing a clear distinction in the transition between the “hard” and the “soft” protein corona. Herein, we introduce the combination of confocal laser scanning microscopy (CLSM) with microfluidics (MFs) as a versatile technique to study protein adsorption onto particles. MF setups provide well-defined systems, as important parameters such as channel dimensions and flow rates are precisely controlled. Such well-defined systems facilitate reproducibility of experimental data and therefore allow standardization of experiments and data reporting.34,35 By monitoring the increase in fluorescence intensity, which results from the adsorption of fluorescently labeled proteins onto the particles, valuable information about formation kinetics and stability of protein coronas can be obtained. Protein adsorption occurs after milliseconds and continues over several minutes until an equilibrium is reached. Using this technique, three different adsorption regimes are observed, which we hypothesize are linked to the formation of three unique phases. Proteins that adhere with high adsorption rates are directly bound to the particle surface and form the first phase, P1hard. This layer is surrounded by irreversibly bound proteins that interact with proteins that have already adsorbed, thereby forming the second phase, P2hard. Both phases belong to the socalled “hard” protein corona and are stable under physiological conditions. A third, loosely attached, and therefore unstable, phase forms the “soft” protein corona (Psoft) as the outer layer. The versatility of this approach is further demonstrated by studying the influence of surface chemistry and applied flow rates on protein adhesion. We demonstrate that low-fouling zwitterionic materials prevent the adsorption of irreversibly bound proteins, but that a “soft” protein corona still forms on them. Additionally, by varying the flow rate, we show that
■
EXPERIMENTAL SECTION
Materials. All chemicals were of analytical grade and used as received without further purification, except for copper(I) bromide (CuBr), which was purified by washing sequentially with glacial acetic acid, absolute ethanol (EtOH), and diethyl ether, followed by drying under vacuum. 2-Hydroxyethyl methacrylate (HEMA) was passed through an aluminum oxide column prior to polymerization to remove the inhibitor. High-purity water (Milli-Q water) with a resistivity of >18.2 MΩ cm was obtained from an inline Millipore RiOs/Origin water purification system (Millipore Corporation, Massachusetts, USA). Device fabrication was carried out using a Sylgard 184 silicone elastomer and the corresponding curing agent from Dow Corning (Michigan, USA). Silicon wafers (diameter 3 in.) were obtained from Si-Mat Silicon Materials (Germany). Developer mr-DEV 600 and photoresists Nano SU-8 50/SU-8 100 were purchased from MicroChem Corporation (Massachusetts, USA). Silica particles with an average size of 16.4 μm [isoelectric point (IEP) 2] were obtained from microParticles GmbH (Germany). Cy3- and Cy5-labeled human serum albumin (HSA) were obtained from Nanocs (Boston, USA) (IEP 5.3). Dulbecco’s phosphate-buffered saline (DPBS) and NaOH pellets (97%) were purchased from Aldrich (Missouri, USA). Silica particle functionalization was carried out using (3-aminopropyl)triethoxysilane (APTES, 98%), ammonia (NH3, 28−30%), pyridine (anhydrous, 99.8%), tetrahydrofuran (THF; anhydrous, 99.9%), and α-bromoisobutyryl bromide (98%), which were all purchased from Aldrich. For the surface-initiated atom transfer radical polymerization [SI-ATRP, HEMA (97%)], N,N,N′,N″,N″-pentamethyldiethylenetriamine (PMDETA, 99%), 2methacryloyloxyethyl phosphorylcholine (MPC), CuBr (98%), and nitric acid (70%) from Aldrich were used. Fluorescence labeling was carried out using AF488-N-hydroxysuccinimide (NHS) purchased from Thermo Fisher (Victoria, Australia) and dimethyl sulfoxide (DMSO) (anhydrous, >99%) that was obtained from Aldrich. Albumin from human serum (lyophilized powder, >96%), purchased from Aldrich, was used for protein corona formation. Fabrication of MF Devices. The MF chips were prepared according to a modified literature method to mimic the in vivo environment.36,37 A linear channel design (for the static in situ studies) and a cross-shaped mixer geometry (for the dynamic in situ studies) were used, both with dimensions of 250 μm × 150 μm (height × width) for the main channel. Structures were designed using the AutoCAD 2013 (Autodesk) software. This network was printed on a photomask foil using a soft photographic emulsion gel. To replicate this design onto a silica wafer, spin-coating cycles of a negative epoxy-based photoresist (SU-8) were applied using a mask aligner (Süss MicroTec). An ∼100 μm thick layer was obtained. Soft baking (65 °C, 10 min) steps and UV patterning were performed after each spin-coating step by placing the mask onto the wafer and exposing it to UV light. Unexposed photoresist was removed using a developer prior to a final hard baking (95 °C, 30 min) step. Soft lithography was performed by pouring a 10:1 w/w mixture of polydimethylsiloxane (PDMS) base and curing agent onto the silicon master. After degassing for 45 min in a desiccator and subsequent drying for 2 h at 75 °C, the replicate was removed from the master, and inlet ports for fluids were added. To obtain a three-dimensional (3D) channel structure, two PDMS devices were sealed together after plasma activation (air plasma, 5 min) and dried overnight at 35 °C. Immobilization of Silica Particles for Static Experiments. The silica particles were immobilized in the channel prior to in situ measurements. The particles were first washed with DPBS via repeated centrifugation (1000g, 1 min, 3×) and resuspension steps and subsequently injected into the channel (5 μL concentrated particle stock solution) with a pipette. The solution of the particle dispersion was then allowed to evaporate overnight at 20 °C, resulting in immobilized silica particles. A particle monolayer was obtained by B
DOI: 10.1021/acsami.8b14307 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces injecting DPBS (10 μL) into the channel to remove excess silica particles. MF−CLSM Static Experiments. In situ experiments for protein corona formation were carried out by combining CLSM with MF. CLSM images were taken using a Leica TCS SP8 equipped with HCX PL FLUOTAR 5×/0.15 DRY, HCX PL FLUOTAR L 20×/0.40 DRY and 40×/0.60 DRY, as well as Fluotar VISIR 25×/0.95 WATER, HC PL APO CS2 63×/1.20 WATER and HC PL APO CS2 63×/1.30 GLYC objectives. The setup was also equipped with 405, 488, 514, 552, and 638 nm lasers, two photomultipliers (PMTs), one transmission PMT (PMTtrans), and an ultrasensitive hybrid detector. Confocal two-dimensional (2D) images were obtained with an ultrafast resonance scanner (12 000 Hz) in the x−y image scan mode. The 638 nm laser was set to 95% with a gain of 500 V for PMT2 and 230 V for PMTtrans. The detection range was set to 660−749 nm. For the measurements, the 3D PDMS MF device was placed on top of a standard glass slide and located under the CLSM microscope. HSA−Cy5 solution in DPBS with a concentration of 0.2 mg mL−1 was injected via syringe pumps (Nemesy, CETONI GmbH) equipped with gas-tight Hamilton syringes. This concentration was chosen on the basis of the fluorescence intensity and the detection limits of the instrument. The devices were connected to the syringes via OriginalPerfusor lines (Type IV-Standard; B. Braun Melsungen AG, Germany) and medical grade polyethylene microtubings (0.38 mm inner diameter × 1.09 mm outer diameter; Scientific Commodities Inc., Arizona, USA). The applied flow rate was 600 μL h−1. CLSM images were taken before the protein solution was injected into the channel, every ∼0.8 s for the first 60 s, and every 30 s from 1.30 to 300 s. From 360 up to 600 s, images were taken every 60 s. A 120 s time interval was applied from 720 to 1200 s. The time interval changed to 300 s between 1500 and 3600 s. The last image was taken after 5400 s. A standard fluorescence curve was constructed as a function of HSA−Cy5 bulk concentration, which showed linear correlation of fluorescence intensity with protein concentration (Figure S1). MF−CLSM Dynamic Experiments. For the dynamic in situ imaging, the same setup as that for the static measurements was used. However, a cross-shaped mixer was used as the particle, and protein solutions were injected simultaneously. Z-scan series were taken in the x−y−z mode and reconstructed by LAS X (Version 2016) software. The focus was adjusted to be at the bottom and top of the MF channel (93.52−85.63 μm) with an increment of 4.71 μm. For the coating experiments, both the 552 and 638 nm lasers were used to detect HSA−Cy3. Both laser intensities were set to 95% with PMT1 = 600 V, PMT2 = 500 V, and PMTtrans = 230 V. Detection ranges were 564−620 nm for HSA−Cy3 and 660−749 nm for HSA−Cy5. The particle dispersion (5 wt % in DPBS) was injected into the middle channel, and the protein solutions (0.2 mg mL−1 in DPBS) were injected into side channels. Flow rates were kept constant for all three channels (20 000 μL h−1 for initial injection, measurements were taken with an applied flow rate of 500 μL h−1 per channel). Particle concentration was measured on a flow cytometer (Apogee Micro Flow cytometer). Protein Corona Stability Measurements. The MF−CLSM static setup was used for the stability measurements. Studies were conducted after protein corona formation (which occurred within 90 min) with HSA−Cy5 on bare silica particles. DPBS was injected into the channel with a constant flow rate of 600 μL h−1 for 5 min following injection of concentrated NaOH solution (17.6 M). CLSM Data Analysis. CLSM images were analyzed using ImageJ processing software. Particle fluorescence intensity was averaged over 10−12 different particles per data set (Figure S2). Particles were randomly chosen, however, with no overlap with each other and close to the area of background measurement. For the intensity measurements, the following assumptions were made: (1) background intensity is constant throughout the channel; (2) background intensity is constant over the area of one particle; (3) photobleaching is negligible, as exposure times are
